To Know Her Is To Love Her: The Truly Personal, Personal Assistant

Her

If you’re interested in where human computer interfaces are going and you haven’t seen Spike Jonze’ latest film, ‘HER’ yet; stop reading this and go to the movies now. It’s a fascinating, cautionary tale of a man who falls in love with his automated personal assistant named Samantha (the voice of Scarlett Johansson).

This is categorically not SIRI. Where SIRI is basically a rules-based natural language search engine, Samantha is an AI-driven learning entity: constantly expanding her knowledge base and creating connections between phrases, ideas, preferences etc. Here’s a quote from the film where Theodore (Joachin Phoenix) has just started using the new software:

Samantha: Well, I take it from your tone that you’re challenging me. Maybe because you’re curious how I work? Do you wanna know how I work?

Theodore: Yeah, actually, how do you work?

Samantha: Well, basically I have intuition. I mean, the DNA of who I am is based on the millions of personalities of all the programmers who wrote me. But what makes me, me is my ability to grow through my experiences. So basically, in every moment I’m evolving, just like you.

Theodore: Wow. That’s really weird.

Samantha: Is that weird? Do you think I’m weird?

[Theodore laughs]
Theodore: Kind of.

Samantha: Why?

Theodore: Well you seem like a person, but you’re just a voice in a computer.

Samantha: I can understand how the limited perspective of an un-artificial mind would perceive it that way. You’ll get used to it.

[Theodore laughs again]
Samantha: Was that funny?

Theodore: Yeah.

Samantha: Oh, good, I’m funny.

While it may be an extreme example, this is not just science fiction. For a while now researchers have been concerned that a rules based approach to natural language processing would eventually reach a dead end. That’s why SIRI makes coy jokes when she (it) doesn’t understand something or simply pulls up a generic web page and puts the onus on the user to find what they want.

In order to move beyond this dead end response, the system has to be able to learn ‘on the fly’ in order to improve both recognition and the results it brings back from any query. It needs to understand context and the only way to do that is to sift through tremendous amounts of data and make connections that are meaningful.

When someone in London tells you, ”I want to go to LA tomorrow”, you instantly know that they mean they want to go to the United States by plane within 24 hours. Furthermore you may know them well enough to understand that they fly on United regularly and what their travel preferences are: including aisle seat, business class, meal preferences, hotels etc.

Who’s Assistant Are You?

So this brings me to my dream world. I want to stop having to conduct business with individual enterprises. I don’t want to contact my bank’s IVR or smartphone interface to get my balance.

If I want to go to LA, I don’t want to contact the airline or the hotel. I want Samantha to do it! She has all my preferences, all my identification and verification details and my history. If I say to her (it) “I want to go to LA tomorrow” She should simply go out to the relevant companies and pull back the information and ask me to confirm the purchases.

Wally: “Samantha, I want to go to LA tomorrow”

Samantha: “Sure, Wally. I can get you on United at 2:30 in seat 12C just like last time. Do you want me to book it?”

Wally: “Yes please”

Samantha: “Should I use the AMEX?”

Wally: “Yup”

Samantha: “Okay and do you want a cab to take you to the airport?”

Wally: “Sure.”

Samantha: “From home?”

Wally: “Great.”

Samantha: “Okay it’ll be downstairs at 12 noon. I’ve put the details and your boarding pass on the phone.”

Wally: “Thanks.”

Samantha: “And don’t forget to pack your umbrella, it’s going to rain in LA all week.”

No more, “Please enter your frequent flyer number.” No more “What’s the expiry date on the card?” Just direct interaction between my system and anything I want to book, buy, query etc. Furthermore the interface travels with me on my phone and all my connected devices. It becomes, quite literally, the persona of my digital world. It can shop, start my car, control my smart home, screen my calls, maybe even do my taxes.

If the airlines and banks and hotels have the APIs available, this shouldn’t be too hard. In fact it’s all the more reason for them to make the data available in the interest of being easy to do business with. This eliminates me having to learn how to use one or another website or IVR ever again. I’ll just leave it to HER.

Wally Brill

One of the founding fathers in voice interface innovation, today Wally Brill heads up the user experience team at VoxGen and remains deeply involved with major brands on their multichannel strategies.

  • Really interesting thoughts and a good software without any doubt!